Structure prediction in Materials Science and Characterisation with EELS in the low-loss regime

Lead Research Organisation: University of Oxford
Department Name: Materials


This project is to enable new developments to the CASTEP Density Functional Theory (DFT) code to provide accurate low loss EELS (Electron Energy Loss Spectra). This necessary to interpret high resolution measurements made possible with the latest generation of Scanning Transmission Electron Microscopes (STEM). This includes the EPSRC National Facility for Aberration-Corrected Scanning Transmission Electron Microscopy (SuperSTEM) located at Daresbury Laboratories. Together, the combination of quantum mechanical DFT calculations and STEM measurements present a powerful tool for materials characterisation. Work in the last decade has shown that it is now possible to use techniques based on quantum mechanics to predict the structure of hitherto unknown materials. To identify the existence of such materials in nature requires the use of experimental techniques such as EELS, which can be compared to the spectra predicted for the proposed structure from DFT simulations. By enhancing our ability to predict low-loss EELS this project will both expand the range of materials systems we can study, and the certainty and confidence with which we can make predictions. The project will enable simulations of a number of materials systems of importance to future materials and the energy sector; including zirconium oxides used in the nuclear industry, fuel cells, Li based batteries and catalysts.

The CASTEP code is developed by academics at the Universities of Cambridge, Oxford, York, Durham and Royal Holloway (London). It is freely available to UK academics and a supported commercial version is marketed by Dassault Systèmes / BIOVIA (with UK HQ in Cambridge). They provide support to industrial users across a wide variety of sectors including: pharmaceuticals, catalysis, energy.

This project delivers new capabilities to accelerate the development of novel materials, particularly for energy applications. However, note that the general nature of quantum-mechanical modelling means that the tools can be applied to a very wide range of materials. In the first instance the beneficiary is the UK company Johnson-Matthey who will supply experimental data on catalysts and fuel cell materials. Through this project they will enhance their understanding of the structure and composition of these materials. However, by working with a leading scientific software company (Dassault Systèmes / BIOVIA) we ensure that the tools developed will be made available across Materials sector - ensuring maximum impact.

The Themes are:

Physical sciences


10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/R512333/1 01/10/2017 30/09/2021
2109852 Studentship EP/R512333/1 01/10/2017 30/09/2021 Xinlei Liu
Description The project can be divided into two parts, improving the current modelling method and applying the method to different materials system studies.

As for the modelling method part, in the density functional theory (DFT) domain, I tested Becke-Johnson, Tran-blaha and Scan which are methods that have never been used to modelling the electron energy loss spectrum (EELS). However, there is no significant improvement of the modelling results comparing to the traditional method using PBE. Beyond DFT, I tried the time-dependent density functional theory using RPA, but still no major improvement is seen in the modelled results. I am working on the GW+BSE method at the moment.

In the application part, I cooperated with the experimental group studying Beryllium samples from the ITER. I have successfully modelled the EEL spectrums for seven candidate compounds and they show some information when comparing to the experimental EELS. However, I figured that there are some problems associated with comparing the modelled and experimental EELS and I am working on it at the moment.
Exploitation Route I should be able to draw conclusions regarding which method is the best to modelling EELS from the modelling techniques studies. And from the application part, take beryllium as an example, I should be able to figure out what corrosion products exist in the beryllium sample which is also helpful to fully understand the corrosion mechanisms of beryllium.
Sectors Electronics,Energy,Manufacturing, including Industrial Biotechology

Description Beryllium from ITER 
Organisation University of Oxford
Department Department of Materials
Country United Kingdom 
Sector Academic/University 
PI Contribution I used the modelling methods to model the electron energy loss spectrums of seven candidate structures of beryllium corrosion products. The modelling spectrums help better interpret the experimental ones.
Collaborator Contribution The experimental data is a good reference of my current modelling method.
Impact A poster has been made to join the MMM Hub conference & user meeting 2019 which took place in London.
Start Year 2019